Scribd Logo
Upload

Welcome to Scribd!

  • 15 views

    Client - Centric Consistency Models

    Uploaded by

    nvrsayno
    Client-centric consistency models focus on consistency for individual clients rather than the entire system. Some client-centric models include: Monotonic reads, which guarantees that a client will see their own successive reads of a data item in the same or a more recent state. Monotonic writes guarantees a client's own writes are seen in the order they were performed. Read-your-writes ensures a client sees the effects of their own writes in subsequent reads. Write-follows-reads means a client's write is based on the most recent value they previously read. These models aim to provide consistency on a per-client basis in systems with mostly reads, lazy updates, and no concurrent writes.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content

    You might also like

    Client - Centric Consistency Models

    Uploaded by

    nvrsayno
    15 views16 pages
    Client-centric consistency models focus on consistency for individual clients rather than the entire system. Some client-centric models include: Monotonic reads, which guarantees that a client will see their own successive reads of a data item in the same or a more recent state. Monotonic writes guarantees a client's own writes are seen in the order they were performed. Read-your-writes ensures a client sees the effects of their own writes in subsequent reads. Write-follows-reads means a client's write is based on the most recent value they previously read. These models aim to provide consistency on a per-client basis in systems with mostly reads, lazy updates, and no concurrent writes.

    Original Description:

    Original Title

    Sanjib CCCP

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    Client-centric consistency models focus on consistency for individual clients rather than the entire system. Some client-centric models include: Monotonic reads, which guarantees that a client will see their own successive reads of a data item in the same or a more recent state. Monotonic writes guarantees a client's own writes are seen in the order they were performed. Read-your-writes ensures a client sees the effects of their own writes in subsequent reads. Write-follows-reads means a client's write is based on the most recent value they previously read. These models aim to provide consistency on a per-client basis in systems with mostly reads, lazy updates, and no concurrent writes.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    Download as pdf or txt
    15 views16 pages

    Client - Centric Consistency Models

    Uploaded by

    nvrsayno
    Client-centric consistency models focus on consistency for individual clients rather than the entire system. Some client-centric models include: Monotonic reads, which guarantees that a client will see their own successive reads of a data item in the same or a more recent state. Monotonic writes guarantees a client's own writes are seen in the order they were performed. Read-your-writes ensures a client sees the effects of their own writes in subsequent reads. Write-follows-reads means a client's write is based on the most recent value they previously read. These models aim to provide consistency on a per-client basis in systems with mostly reads, lazy updates, and no concurrent writes.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    Download as pdf or txt
    of 16

    Client-Centric Consistency Models

    Clients view of data-store


    Ideally: black box -- complete transparency over how data is stored and managed

    Introduction
    n

    The models weve talked about have system wide consistency for systems where concurrent processes may be updating the data store. Now, lets assume that there arent simultaneous updates to the data store, or that when they happen they can be easily resolved, and that most operations involve reading data.

    Introduction (contd)
    n

    Consider the Domain Name System


    In these systems, the vast majority of the operations are reads, and essentially all writes are done by central authorities (web page owners, host masters), so we never have write-write conflicts. n The only situation that needs to be handled are read write conflict. n Update propagate in a lazy fashion.
    n

    Eventual Consistency
    If no updates take place for a long enough period time, all replicas will gradually (i.e., eventually) become consistent.
    n

    Situations where eventual consistency models may make sense


    Mostly read-only workloads n No concurrent updates
    n

    Basic Architecture

    n n

    How well does eventually consistent (EC) work for mobile clients? Client-centric is for this. Consistent for a single client.

    Client-Centric Consistency Models


    n

    Different Client-centric consistency models


    Monotonic reads n Monotonic writes n Read-your-writes n Write-follows-reads
    n

    Goal: Show how we can perhaps avoid system wide consistency by concentrating on what specific clients want, instead of what should be maintained by servers.

    Client-Centric Consistency Models


    n

    We describe client-centric consistency models with the following notation


    xi[t] is the version of data item x at local copy Li at time t n WS(xi[t]) is the set of write operations at Li that have taken place since initialization (resulting in xi[t]) n If operations in WS(xi[t1]) have also been performed at local copy Lj at a later time t2, we have WS(xi[t1];xj[t2])
    n

    Monotonic Read Consistency


    n

    If a process reads the value of a data item x, any successive read operation on x by that process will always return that same value or a more recent value. Sequence (a) satisfies monotonic read consistency, but sequence (b) does not.

    Example
    n

    distributed e-mail database


    Mails can be inserted at different locations n Changes are propagated lazy. n A user reads his mail in Kolkata. n Assume that he does not change the mailbox, only reads the mails n He flies to New York and opens his mailbox n Monotonic read consistency guarantees that he sees all mails in his mailbox, he saw in Kolkata.
    n

    Monotonic Write Consistency


    n

    A write operation by a process on a data item x is completed before any successive write operation on x by the same process (all write operations are propagated to all replicas in order). Sequence (a) satisfies monotonic write consistency, but sequence (b) does not

    Read-Your-Writes Consistency
    n

    The effect of a write operation by a process on data item x will always be seen by a successive read operation on x by the same process. Examples for missing this consistency
    We change a web page but we still get the old version from the cache we have to push reload n We change the password but the propagation takes
    n

    Graphical Representation
    n

    Sequence (a) satisfies read-your-writes consistency, but sequence (b) does not

    Writes-Follow-Reads Consistency
    n

    A write operation by a process on a data item x following a previous read operation on x by the same process is guaranteed to take place on the same or a more recent value of x than that which was read. Examples for missing this consistency
    n

    In a network discussion group.

    Graphical Representation
    n

    Sequence (a) satisfies read-your-writes consistency, but sequence (b) does not

    Thank You

    You might also like